ITHome · Published 2026-09-10
DeepSeek V4.1 Flash Model Goes Live on the National Supercomputing Internet
Summary: DeepSeek V4.1 Flash went live today on the National Supercomputing Internet. It is a 552B-parameter MoE model built on a brand-new Causal-Encoder-Decoder architecture that significantly cuts costs. Its benchmark scores surpass those of the flagship model, and its greatly compressed KV Cache effectively lowers Agent usage costs. #AILargeModels
01THE STORY
About this story
Background and context
Read the linked page for the details, and follow its citations for further context.
Read claims in the context of the original page. Claims remain attributed to their sources. Collection is not proof of the claims or practical outcomes.
Reading progress is saved in this browser. Save or add a note →
Looking back over time
02TIMELINE
Looking back over time
- First discoveredFound this reading link via ITHome.
Collection and archive details
Record location: live register
First discovered: 2026-09-10; added to the library: 2026-09-11 (Beijing time).
Collection and source publication have distinct dates. A library addition does not establish a daily selection; retained selection dates are listed above.
03SOURCES
Original links and related material
- Read the originalSourceITHome
Sources and evidence · 2 links
Each link has a purpose and scope. Link counts do not establish independent confirmation; direct support applies only to the statements specified below.
This page keeps the original link and a short account. Follow the links below for the publisher’s full explanation.
- ITHome articleDetails not yet added
The specific statements supported by this link have not yet been documented.
- ITHome feedDetails not yet added
The specific statements supported by this link have not yet been documented.